I made a quick test connection with BC639 transistor.
I used a 12vdc led lamp with built-in resistors and such.
When connected directly to Vin and GND (I use 12v power for my UNO) it works as intended.
BUt when I connect it to transistor it only lights 2 out of 5 leds. When I measure voltage over the lamp it gives 7,5V
What am I doing wrong? Code-wise I used Blink.
BUt when I connect it to transistor it only lights 2 out of 5 leds.
You loose between 0.7 to 1.2V across the emitter / collector junction.
Otherwise if you have 5 LEDs in seriese does each one have a seriese resistor? You only need one resistor for seriese LEDs otherwise each one drops a voltage and that adds up.
Well the 5 leds thing is a black box kind of thing. I can't get inside it and I know it lights all 5 correctly when I apply 12 volts to it straight. It's a work light for any machine that uses 12vdc.
BC635-D.PDF (92.9 KB)
So have you wired it up to a transistor when you see the problem. The with a wire link short out the collector / emitter. If it then works you have either
Grumpy_Mike:
So have you wired it up to a transistor when you see the problem. The with a wire link short out the collector / emitter. If it then works you have either
- too small a base resistor, what value are you using?
- a faulty transistor
- you need another 2V , that is a 14V supply
- you should use a logic level FET instead of a transistor.
Depending on the load current 1K might be too high, try 270R, or look at the transistor’s data sheet for the saturation base current.
Further, Looking at the data sheet attached further up, I don’t think that transistor is good to 1A. According to the graphs It’s hfe (gain) falls off a cliff well before that.
Also some voltage may be "lost" in the breadboard and jumper wires. I am using poor China stuff but if I used the setup as the image shows I would have "high" resistance in the wires - several Ohms. Since 1 Ohm = 1 V voltage drop at 1 A of current!
A BC639 has a low current current gain of 40-180.... but to obtain it's saturation Vce of 0.5v needs 50mA base current with a collector load of 500mA. The 1k base drive resistor you show only provides about 4mA
Bipolar transistors aren't very good as high current switches. Try eg a FQP30N06L logic level MOSFET .
Allan
davetcc:
I think I see a 1 amp load there. If that is right then let’s assume gain is 100...
Common misunderstanding.
Transistor gain is irrelevant for switching (when collector voltage drops below base voltage).
The BC639 datasheet states a Ib:Ic of 1:10.
Leo..
The gain value hfe is for when the base-collector junction is reverse-biased, ie the collector is at least a volt
or two above the emitter.
In hard saturation the collector can be 0.1V above the emitter, and the base-collector junction is forward-biased.
the factor of 5 to 20 "gain" in saturation is due to the law of mass-action and the emitter being doped about
1000 times more strongly than the collector.
Typically hard saturation only matters for high power, where the difference between Vce being 0.7V and 0.1V is
a factor of 7 in heat dissipation in the switching device - you are actually trading off reduced heat dissipation
against extra power needed to provide current to the base.
MarkT:
Typically hard saturation only matters for high power...
Which in this case might be important.
If the LED light is using 500mA, then 1volt across the transistor is a junction temp rise of 100C above ambient.
Leo..
davetcc:
I think I see a 1 amp load there. If that is right then let’s assume gain is 100.1000ma / 100 = 10mA
Allowing double would be 20mA.Then on the base we get
5V / 1000ohm gives 5mA
This would leave you short of base current. You would need a 220 ohm resistor there but that is cutting it fine on the Arduino side. Is this transistor able to pass 1a emitter to collector? For 1A I’d consider a MOSFET or Darlington buffer.
ANY transistor circuit design that depends on Hfe being a certain value is a bad design.
For any transistor you use, you should assume that the Hfe is very low... maybe 1/10 of what the spec sheet says. If that doesn't give you enough, then you need a Darlington pair (or a MOSFET if you're switching).
The gain of a linear transistor amplifier should be determined by the bias value, the collector and emitter resistor values, without regard to Hfe.
Switched the transistor to a https://www.vishay.com/docs/91015/sihf510.pdf
Switched the 1k ohm to 270 ohm.
To no avail.
An IRF510 is NOT a logic level mosfet, and might or might not work.
FIRST measure the current draw of that light, THEN decide what transistor you need.
Leo..
I was cheated at the component store...
Lamp takes 700 mA
Well well, dumdum strikes again. After some surfin' and turfin' I found out how to actually calculate and in the end a 120 ohm resistor on the base of the BC639 gave enough to light up the whole shebang. It measures now at 10V, so maybe the resistor could be taken out all together?
Since the BC639 has hfe of 40-160 and lamp is 700mA. On the info I found on the interwebs it says Ic/hfe=Ib I got (700mA/40)= 17,5mA. Then they recommended to double that. (35mA) and then (5V-0,6V)/0,035A=125.7ohms. So i took a 120 ohm.
Now I am trying to figure how the calculations go for a fet. (IRF510 is what I have right now)
It says on the datasheet, that the threshold voltage is 2-4V on IRF510. if Id should be 0,7A and IRF has transconductance of 1.3 (seems a bit low? Am I looking at the wrong place).
tuomasjar:
After some surfin' and turfin' I found out how to actually calculate......so maybe the resistor could be taken out all together?
Since the BC639 has hfe of 40-160 and lamp is 700mA. On the info I found on the interwebs it says Ic/hfe=Ib I got (700mA/40)= 17,5mA. Then they recommended to double that. (35mA) and then (5V-0,6V)/0,035A=125.7ohms. So i took a 120 ohm.
Now I am trying to figure how the calculations go for a fet. (IRF510 is what I have right now)
You don't calculate the base current of a transistor switch with Hfe.
Nooooo, then the Arduino pin would be shorted to ground through a diode (the BE junction).
Hfe is irrelevant for a switch (see post#10). You need ideally 70mA (1/10 of Ic) in the base.
An Arduino pin can't deliver that.
You could drop that to 20mA, to keep the Arduino happy, but that won't fully saturate the transistor.
CE volt drop could be 2volt with 20mA base current (I think you already measured that).
That will dissipate 2volt*.07A= 1.4watt in the BC639. Enough to raise the internal temp by 280C. Crackle/pop.
The only way to switch 700mA is with a darlington (e.g. TIP120), or with a logic level mosfet.
A mosfet is better, because of the potentially lower volt drop.
Leo..
tuomasjar:
Now I am trying to figure how the calculations go for a fet. (IRF510 is what I have right now)
You don't calculate that either. Just put a 10volt logic signal on the gate.
Oh, wait. The Arduino can't deliver that either.
Leo..
Wawa:
You don't calculate the base current of a transistor switch with Hfe.Nooooo, then the Arduino pin would be shorted to ground through a diode (the BE junction).
Hfe is irrelevant for a switch (see post#10). You need ideally 70mA (1/10 of Ic) in the base.
An Arduino pin can't deliver that.
You could drop that to 20mA, to keep the Arduino happy, but that won't fully saturate the transistor.
CE volt drop could be 2volt with 20mA base current (I think you already measured that).
That will dissipate 2volt*.07A= 1.4watt in the BC639. Enough to raise the internal temp by 280C. Crackle/pop.The only way to switch 700mA is with a darlington (e.g. TIP120), or with a logic level mosfet.
A mosfet is better, because of the potentially lower volt drop.
Leo..
Right-o. Just gotta walk to FLCS and purchase a bunch of tip121 (at least those I could find on the website)
How about running a pwm signal through darlington? or do I need a mosfet for that?
